Japanese Kokai Patent Application No. Hei 11[1999]-160335 describes a conventional image velocity computer. The JP '335 publication teaches that, in order to compute the image velocity, an optical flow (i.e., a vector indicating movement on an image) of a target object is computed. The image velocity is computed based on the direction and the level of the computed optical flow.
To calculate the optical flow, a gradient computation method is applied to an image that is generated by an on-vehicle camera. The gradient computation method assumes that the optical flow can be computed by solving simultaneous constraint equations based on discretely sampled images. The method also assumes that the gray values of the target object are retained. However, the gray value distribution of the target object fluctuates significantly when the target object is moving rapidly. For this reason, it is difficult to accurately compute the image velocity when the target object is moving rapidly. Accordingly, a need exists for an improved image velocity computer that can accurately compute image velocity even when an image velocity computation target is moving at a high velocity.